Current Meeting Report

2.2.1 Operation of the IESG/IAB Nominating and Recall Committees (nomcom)

NOTE: This charter is a snapshot of the 54th IETF Meeting in Yokohama, Japan. It may now be out-of-date.

Last Modifield: 06/14/2002

G Kullgren <>
General Area Director(s):
Harald Alvestrand <>
General Area Advisor:
Harald Alvestrand <>
Mailing Lists:
General Discussion:
To Subscribe:
In Body: subscribe
Description of Working Group:
RFC2727 (BCP10) specifies the process by which the members of the IAB and IESG are selected, confirmed, and recalled. It has been two years and three nominating committees since the document was updated. The experience has identified a number of areas where the process could be both improved by additional clarifications and updated with additional guidance.

This working group will consider all such clarifications and suggestions and produce a revision to RFC2727. However, the basic mechanism of selection, confirmation, and recall will not be changed by this group. It is only concerned with the details of the current process.

The input document will be draft-galvin-rfc2727bis-00.txt.

Goals and Milestones:
MAY 02  Initial set of issues to be addressed selected
JUN 02  Revised I-D published
JUL 02  Revised I-D published, WG Last Call started
AUG 02  Submission to IESG for publication
  • - draft-ietf-nomcom-rfc2727bis-01.txt
  • No Request For Comments

    Current Meeting Report

    The Nomcom WG meeting was chaired by Avri Doria and minutes were recorded by Robert Elz and by Ron Bonica.

    The agenda was reviewed and there were no changes requested.

    The WG was reminded that the charter and the milestones mandated this WG to be short lived, I.e. this should be the only WG meeting at an IETF (in the best of all possible worlds).

    Avri, acting as a proxy for Jim Galvin, reported on the changes to RFC 2727. These are listed in the draft itself (section 7) and in the meeting slides.

    Several of the changes prompted the meeting discussions outlined below. These discussions are referred back to the design team.

    - On the issue of the IAB/IESG deciding whether to fill a vacant position.

    Q: What is a mid-term vacancy

    A: anything after Nomcom starts

    Q: That means IAB/IESG can decide that Nomcom isn't liked, and just hold over a vacancy that occurs early in Nomcom process for the following Nomcom.

    A: Yes, could happen.

    Q: Maybe need division into year halves. If a vacancy occurs early in year current Nomcom decides, later in year, either the next Nomcom must handle or the IAB/IESG has discretion.

    A: Perhaps the draft should include advice to that effect

    - Robert Elz: On #8: much of what is in rule 7 (changes) would be better in rule 8 - so that what a nomination must contain is in the section about making a nomination, not in the one about sending a request for nominations to be made.

    Avri: will take back to editor.

    - #12 - On the issue of archives:

    Eric Gray: Issue with that is - chair is supposed to be able to provide info for the next Nomcom, is the chair intended to be part of that.

    A: the archive goes away, chair keeps mental records.

    The next agenda item was a discussion of Open issues, i.e. issues for which no changes were yet made to the draft.

    - Length of Process & Timing This has been discussed a lot on the list and was discussed further during the meeting.

    Robert Elz (kre): Spoke to Brian Carpenter's issues, as understood from his email. It should be possible for the Nomcom to finish before Christmas, to avoid being slowed down by that interruption, hence start earlier. Then confirming body can operate in late Jan/Feb.

    Christian Huitema: agrees with this, too much pressure on confirming body to simply agree otherwise process fails, but disagrees with kre's interpretation of Brian's point.

    Bob Hinden: wants to give confirming body enough time, but not so much that they have time to reject candidates because they don't like them, so don't want too much time for confirming body - Nomcom should make the decisions, it should be rare for confirming body to push back, except where a process problem is identified.

    Harald Alverstrand: - giving confirming bodies more time is good, giving Nomcom itself more time is bad. His experience - pressure on Nomcom & amount of time it takes increases over time, adding more time will just give Nomcom more time to use up. Don't increase time available for Nomcom itself.

    Fred Baker: don't limit Nomcom's time. They are already working fast. Nomcom has had to deal with having too little time, if it is constituted too close to 3rd IETF, it has no time to prepare.

    Avri: current draft attempts to allow for enough time.

    Ron Bonica: should be specific about how much time the confirming body has

    Avri: isn't that in there already?

    kre: hard to figure out from the draft - info might all be there but very hard to find.

    Avri: yes, probably need to make that more clear

    Eric: agrees with Fred - timing of IETF meetings themselves isn't fixed - need to base timings on 1st IETF (end of Nomcom) not the 3rd IETF, as that makes Nomcom duration random

    Avri: OK.

    Avri: Time period - 2 months for Nomcom to work, is that OK? There was no dissent on this point.

    Time periods need to be more explicit. This is to be passed on to the editor.

    Harald A: time that Nomcom is active should include at least one IETF meeting needs to be made explicit

    Avri: that can probably also be concluded from the draft now, just isn't explicit - more explicit times are needed.

    - The next issue to be discussed was that of accountability - of the Nomcom to community &/or confirming bodies.

    The question is does more need to be said about this in the draft? What does accountability mean?

    Christian Huitema: It is part of the process, like picking a jury.

    Avri: Does liaison responsibility have any effect

    Christian: Do you mean can they be sued?

    Avri: No... Is there an assumption that the liaisons would let the community know if Nomcom is going off the rails.

    Christian: can't cover situations like this

    Steve Hanna: Expect of Nomcom what we expect of juries - by taking cross section, we expect that we will get people who have common sense, will represent a reasonable cross section of what typical IETF members might want. Sometimes might get non-representative grouping - perhaps some members who are a bit cracked, but perhaps that is representative... This is probably the best we can achieve. More accountability is likely to discourage people from wanting to serve. Unlike jury, Nomcom is volunteer, not imposed, need to be able to attract people.

    Avri: Asked if the general consensus of room is that accountability as it is in draft is fine. No dissent from the room.

    - The next issue to be discussed was whether the names of nominees should be published? This was originally proposed as a means for confirming body to have an idea what was coming, an issue that has been handled by fixing the timing. In meantime this issue has gained a life of its own.

    The current proposal for publishing names was made by Harald on the WG list.

    Bala Rajagopalan: thinks names should be published for a different reason. Need people to have some idea that process is fair - how Nomcom works isn't visible outside community. More than just names, the questionnaire filled in by nominees should be published as well. Need to de-mystify the process.

    Avri asked if recommendation was that all nominees whether accepted or not should be published.

    Answer: No only those who agree to go forward, should be published, and their views.

    Avri pointed out that the questionnaire is not a mandated part of the process and it is up to the Nomcom whether questionnaireds are used or not.

    Joseph Hui - should publish names, shouldn't want privacy

    Eric - should have at least first right of refusal, keep name off the list if not interested. An issue with questionnaires - people tend to be more or less careful/honest depending upon the expected audience. If comments are to be public, then all will be similar, and depend upon what company lawyers allow them to say. Confidentiality helps.

    Harald A: Commented that anyone who expects to run should dare to show face in public - in several cases, positions filled by arm twisting, convince people who didn't really want the job. Would not like to see process where unable to get best people to serve. Wouldn't trust anyone who wants his job.

    Leslie Daigle: supported idea in plenary last time - for purpose of ensuring max input to nominating committee. But concerned at notion of publishing questionnaire. This turns process into an election. Will result in campaigning.

    Bob Hinden: agree with both previous people - comfortable with publishing names of people who have agreed to run, but not questionnaires.

    Christian H: The risk is that selection tends to be old boys club. But process has protection - Nomcom selection process - that avoids this problem. The past nominations contained some surprises - shows this. Point that Leslie made is more important - by publishing names, we are trying to take control of Nomcom process, to make it an election. Notion of who is a candidate is very fuzzy, Fred Baker used to say the reason he was a candidate was that he didn't say no loudly enough,. If list published, then important part will be who is not on the list - and questions when people are not on list, what the reason is that they declined. Should not publish names.

    Bala - if person is selected, then should publish their responses to what they would do. Have more people than are needed, that's why Nomcom exists.

    Mike St Johns: One of the problems, is that some people refuse to oppose others; 'if X is running, I won't run against them.' In the past, as chair, had to apply pressure to get people to be considered, especially when against a strong incumbent who wanted to continue. But is not strongly opposed to publishing list that simply has everyone who was ever suggested. I.e. Harald's proposal.

    Fred Baker: responded to earlier comment questioned amount of arm twisting. Huge amount of arm twisting is required - have to think about the amount of work being required. Fred's experience - in 95 was working in multiple areas, with technical basis - when asked if he would be IETF chair. Fred had to drop lots of technical work which changed Fred's career track. There may be no one actually willing to make a change like that. Not many qualified people, looking for management people from a community of engineers. Also have other criteria - want people from Asia/Europe, have phone calls at odd times of the night, fly more than requires.

    Eric - don't want people who don't want the job, or who don't think they are qualified, to be on list without their consent

    Ron Bonica - publish names of people who have accepted to be candidates. Also agree with Leslie - don't publish Q's.

    Bala: - response to Fred - after someone has agreed to run, then want to know more about the people who are being considered (after selected).

    Dave Crocker: Has been tracking topic closely - doesn't know what problem is being solved. Usually make changes to add features or fix bugs - fears that these changes are based upon abstract notions of fair process. Have been times that people would have liked to know the names, as clearly they didn't get enough feedback. That kind of extra feedback would be better late in the process, publish only smaller list if any.

    Christian: if we do anything, must know why. So far just two reasons. So people can audit that Nomcom is looking at wide enough set of candidates, already have protection for that. Other for people to provide ongoing input - e.g. 'don't want that asshole to be included' - and we don't want that kind of input, that would be not relying on Nomcom, but 2nd guessing.

    Harald: what problem is solved? Being blunt - one problem is that current Nomcom process divides IETF into 2 groups, one of about 100 are asked to state their opinion, and have a fair idea who candidates are. The rest, about 6000, who have no idea at all - would like to move a little away from that. Problem is that need to make a small amount of change so the rest of the people get considered.

    Leslie: problem being solved is ensuring appropriate input to Nomcom, proposed implementation is publishing list, which might not be the best way. Shouldn't be trying to 2nd guess the Nomcom.

    kre: When nominees announced - info on those is given, so don't really need announcement of qualifications of selected candidates, more than is already provided for. On getting extra input for Nomcom, my experience as outsider has been that nominees (once confirmed) are either people that were obviously one of the possibilities, or people I had never heard of, and so couldn't provide input about, even if their name had been announced.

    Avri asked for hum to determine if status quo should be maintained or if names should be published (in some form) - with a process to be defined.

    The humming sounded equal. No consensus for change at this point. The discussion should be continued on the list.

    - Next issue discussed was whether anything further needed to be included on the role of liaisons

    kre: should say that liaison should be bi-directional - should make it clear that liaison should report to community.

    Thomas Narten: should write down what liaisons do, have a fair amount of running code now, but the doc says nothing.

    Avri asked if this should this be in the form of rules or as oral tradition.

    Thomas: should be rules. Not just that previous ones have done/not done.

    General room consensus that doc should say more.

    - The next issue involved controversial decisions

    Avri: Nomcoms know when they are doing something controversial and the Nomcom is now being told they should be ready to back up their decisions.

    Concerning other issues from list.

    Does the room think that the oral tradition section is a good section, i.e. is it good to list practice without rules? And should anything else be covered?

    Eric Gray: Be a good idea to spell out what the procedures approve of. Controversy in 4 categories - not clear which, expect it to be controversial and it is, and it isn't, and expect it not to be and it isn't, or is. Has seen some more controversy than expected, and vice versa. Also can be surprised about what kind of proof of process is required.

    Avri: some of this covered by dialog with confirming body, and can also be covered if liaisons have more to do.

    Fred Baker: On what he believes is the place of rules. Tend to armchair quarterback - make a rule that bad things can't happen. That's hard to do. Don't want to go there. Want to talk about expected goals and outcomes. Where process clarification is required, set that down - but in general, want to depend upon the people appointed to sue good judgment, can't predict what will be required.

    Avri: interpret that as desire for more oral tradition and fewer rules. Asked if that was that the general opinion.

    The room indicated agreement.

    Avri asked if anything else should be added to oral tradition section. There was no response.

    The meeting was then opened to other issues participants might have.

    kre: brought up a definitional issue concerning 'usual definitions' of non-voting members and what that meant in relation to items where the text appeared to be giving them a vote.

    Avri: thinks that all members should vote on all issues of Nomcom practice.

    Bob Hinden: only voting members set the rules, and vote

    Thomas Narten: agrees - only voting members vote, except where explicit.

    Avri: should clarify when non-voting members get to vote

    kre: no, it isn't when that need to be made explicit, but how.

    Issue needs to be reviewed by the list and by the design team.

    Milestones Review:

    We probably won't make July 02 for revised ID & WG last call, but August might be possible. Won't make Aug 2 for submission to IESG, but should be done before next IETF meeting.

    kre: will this mean that this draft isn't ready for next Nomcom, will be well underway by then - have to either tell next Nomcom to follow an I-D, or publish earlier than looks possible, or leave these changes for the following Nomcom.

    Avri asked the room if we should we tell powers that be to follow process outlined in the current draft - especially with regard to timing.

    The room agreed with this.

    There was comment that the new Nomcom should follow the drafted procedures to the extent that they don't contradict existing doc.

    Avri to Harald (in his general area AD capacity): how do we do that? Is fact you have heard it enough?

    Fred Baker: you might want to tell chair of Internet Society.

    Avri: Good idea. Who is that?

    Fred raises his hand.